INDEX

Explanations

phrases beginning with a specific word

oai_token-act-pair · gpt-3.5-turbo

instances of the word "at" indicating specific locations or times

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gravity

-0.78

FTWARE

-0.76

alpha

-0.72

fill

-0.68

Guest

-0.63

Pac

-0.60

istar

-0.60

Lens

-0.60

Russ

-0.59

REDACTED

-0.59

POSITIVE LOGITS

 least

1.25

onement

0.96

 halftime

0.96

abase

0.89

hens

0.85

roph

0.82

 dusk

0.76

rial

0.74

 variance

0.73

 sunset

0.70

Activations Density 0.242%

phrases beginning with a specific word

instances of the word "at" indicating specific locations or times

No Comments

No Known Activations