INDEX

Explanations

mentions of different languages

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "ones"

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

anooga

-0.73

NRS

-0.72

lished

-0.72

ERAL

-0.70

MER

-0.66

yrinth

-0.65

Seattle

-0.65

 liability

-0.65

ancouver

-0.65

MED

-0.64

POSITIVE LOGITS

ones

1.31

anguage

0.90

creen

0.90

uch

0.87

yton

0.85

xus

0.84

ynthesis

0.82

hots

0.81

soever

0.80

omething

0.74

Activations Density 0.006%

mentions of different languages

occurrences of the word "ones"

No Comments

No Known Activations