File size: 4,865 Bytes
3d3ae36
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
{
  "task": {
    "id": "case_21",
    "name": "unique_token_extractor",
    "summary": "Extract unique tokens from a string",
    "examples": [
      {
        "input": [
          "BOS",
          "b",
          "b",
          "a",
          "a",
          "b",
          "b",
          "a",
          "a",
          "a"
        ],
        "output": [
          "BOS",
          "b",
          null,
          "a",
          null,
          null,
          null,
          null,
          null,
          null
        ]
      },
      {
        "input": [
          "BOS",
          "c",
          "c",
          "a",
          "c",
          "a",
          "c",
          "c",
          "a",
          "a"
        ],
        "output": [
          "BOS",
          "c",
          null,
          "a",
          null,
          null,
          null,
          null,
          null,
          null
        ]
      },
      {
        "input": [
          "BOS",
          "c",
          "a",
          "c",
          "c",
          "a",
          "a",
          "c",
          "b",
          "c"
        ],
        "output": [
          "BOS",
          "c",
          "a",
          null,
          null,
          null,
          null,
          null,
          "b",
          null
        ]
      },
      {
        "input": [
          "BOS",
          "c",
          "c",
          "c",
          "b",
          "c",
          "b",
          "b",
          "c",
          "b"
        ],
        "output": [
          "BOS",
          "c",
          null,
          null,
          "b",
          null,
          null,
          null,
          null,
          null
        ]
      },
      {
        "input": [
          "BOS",
          "c",
          "c",
          "c",
          "a",
          "c",
          "b",
          "a",
          "b",
          "b"
        ],
        "output": [
          "BOS",
          "c",
          null,
          null,
          "a",
          null,
          "b",
          null,
          null,
          null
        ]
      }
    ],
    "is_categorical": true
  },
  "program": "tokens_with_indices = rasp.SequenceMap(lambda x, i: (x, i), sop, rasp.indices)\noccurrences_selector = rasp.Select(tokens_with_indices, tokens_with_indices, lambda k, q: k[0] == q[0] and k[1] <= q[1])\noccurrences_per_row = rasp.SelectorWidth(occurrences_selector)\nfirst_occurrences_indices = rasp.SequenceMap(lambda count, i: i if count == 1 else -1, occurrences_per_row, rasp.indices)\ntokens_selector = rasp.Select(first_occurrences_indices, rasp.indices, rasp.Comparison.EQ).named(\"unique_tokens_selector\")\nreturn rasp.Aggregate(tokens_selector, sop, default=None).named(\"unique_tokens\")",
  "components": [
    {
      "id": "L0_MLP",
      "hook": "blocks.0.mlp.hook_post",
      "role": {
        "tag": "COMBINER",
        "note": "Creates a combined 'token + position' representation so later steps can tell which token occurred where."
      },
      "rasp_vars": [
        "tokens_with_indices"
      ],
      "labels": [
        "sequence_map_6"
      ]
    },
    {
      "id": "L1H0_ATTN",
      "hook": "blocks.1.attn.hook_result[0]",
      "role": {
        "tag": "AGGREGATOR",
        "note": "Aggregates matching positions (same token, earlier-or-equal index) into a per-position count of how many times each token has appeared up to and including the current position."
      },
      "rasp_vars": [
        "occurrences_selector"
      ],
      "labels": [
        "selector_width_4_selector_width_attn_output"
      ]
    },
    {
      "id": "L1_MLP",
      "hook": "blocks.1.mlp.hook_post",
      "role": {
        "tag": "MAPPER",
        "note": "Converts the per-position summary into the explicit running count of that token's occurrences so far. (how many times this token has appeared up to here)."
      },
      "rasp_vars": [
        "occurrences_per_row"
      ],
      "labels": [
        "selector_width_4",
        "selector_width_4_selector_width_attn_output"
      ]
    },
    {
      "id": "L2_MLP",
      "hook": "blocks.2.mlp.hook_post",
      "role": {
        "tag": "COMBINER",
        "note": "Computes a per-position routing key that marks the first occurrence positions and marks all others as an invalid index."
      },
      "rasp_vars": [
        "first_occurrences_indices"
      ],
      "labels": [
        "sequence_map_3"
      ]
    },
    {
      "id": "L3H0_ATTN",
      "hook": "blocks.3.attn.hook_result[0]",
      "role": {
        "tag": "AGGREGATOR",
        "note": "Attends to the position (if any) whose 'first occurrences indices' equals the current index, and aggregates the token value from there, or None if no such position exists."
      },
      "rasp_vars": [
        "unique_tokens"
      ],
      "labels": [
        "unique_tokens_1"
      ]
    }
  ]
}