INDEX

Explanations

commands or suggestions advising against certain actions

oai_token-act-pair · gpt-3.5-turbo

the repetition of the word "refrain" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ammy

-0.96

swick

-0.69

NX

-0.69

onomy

-0.62

 Patriarch

-0.62

oÄŁ

-0.62

agency

-0.61

immer

-0.61

NetMessage

-0.60

Hub

-0.59

POSITIVE LOGITS

rences

1.00

 refrain

0.92

 abst

0.85

ministic

0.75

ftime

0.71

untarily

0.70

 answering

0.68

 fools

0.66

ively

0.66

 acknow

0.66

Activations Density 0.012%

commands or suggestions advising against certain actions

the repetition of the word "refrain" in various contexts

No Comments

No Known Activations