INDEX

Explanations

statements describing what actions or decisions should be taken

oai_token-act-pair · gpt-3.5-turbo

statements indicating a necessity or requirement to be fulfilled

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Wid

-0.72

 maze

-0.67

 spectrum

-0.65

plex

-0.64

 stakes

-0.63

 Bomber

-0.62

 Puzzle

-0.62

Haz

-0.61

 Wick

-0.61

Or

-0.60

POSITIVE LOGITS

 judged

1.01

 able

1.01

fitting

0.96

leeve

0.95

acons

0.95

 regarded

0.95

 treated

0.93

arers

0.92

heading

0.90

 considered

0.88

Activations Density 0.088%

statements describing what actions or decisions should be taken

statements indicating a necessity or requirement to be fulfilled

No Comments

No Known Activations