INDEX

Explanations

scientific references, such as citations and figures

oai_token-act-pair · gpt-3.5-turbo

references to figures or data representations in scientific texts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 surround

-0.80

 evict

-0.77

 estate

-0.76

 pronoun

-0.75

 renting

-0.75

 piping

-0.74

 taxpayer

-0.73

 handlers

-0.73

 broadcasting

-0.72

 imperson

-0.72

POSITIVE LOGITS

see

1.55

Figure

1.50

reviewed

1.49

Fig

1.44

Supplement

1.35

Table

1.33

fig

1.22

cf

1.20

 Figure

1.17

particularly

1.17

Activations Density 0.090%

scientific references, such as citations and figures

references to figures or data representations in scientific texts

No Comments

No Known Activations