INDEX

Explanations

positive and negative qualities or characteristics

oai_token-act-pair · gpt-3.5-turbo

phrases highlighting positive aspects or "good things" about a topic

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

chair

-0.82

urated

-0.79

igate

-0.73

mens

-0.73

ALK

-0.72

utm

-0.71

rained

-0.70

-0.69

eyes

-0.69

alf

-0.69

POSITIVE LOGITS

 kicker

0.82

 bonus

0.73

 Bonus

0.70

 Solitaire

0.69

 downside

0.69

 happens

0.64

Bonus

0.64

:]

0.64

 surprises

0.63

 Pinball

0.62

Activations Density 0.197%

positive and negative qualities or characteristics

phrases highlighting positive aspects or "good things" about a topic

No Comments

No Known Activations