INDEX

Explanations

numbers related to measurements or statistics

oai_token-act-pair · gpt-3.5-turbo

numerical values, particularly percentages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

gets

-0.87

sworth

-0.83

lings

-0.79

lihood

-0.78

folk

-0.75

holders

-0.72

metry

-0.71

holder

-0.70

held

-0.69

hold

-0.68

POSITIVE LOGITS

ologue

0.91

%:

0.91

0.87

0.86

mph

0.85

0.83

ocal

0.81

0.80

Activations Density 0.026%

numbers related to measurements or statistics

numerical values, particularly percentages

No Comments

No Known Activations