INDEX

Explanations

instances where correctness or accuracy is being evaluated or affirmed

oai_token-act-pair · gpt-3.5-turbo

statements asserting correctness or validation

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

IFrame

Link

Negative Logits

aden

-0.80

CHO

-0.71

agine

-0.70

EMOTE

-0.69

Das

-0.68

GGGGGGGG

-0.67

 Valhalla

-0.66

atos

-0.66

cheon

-0.65

thin

-0.65

POSITIVE LOGITS

ives

1.00

eous

0.95

ible

0.84

 answers

0.84

 spelling

0.82

ibly

0.81

 guiActiveUn

0.78

yt

0.78

able

0.78

 guiIcon

0.77

Activations Density 0.016%

instances where correctness or accuracy is being evaluated or affirmed

statements asserting correctness or validation

No Known Activations