INDEX

Explanations

information related to numbers, especially rankings or statistics

oai_token-act-pair · gpt-3.5-turbo

phrases related to numerical data and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

stage

-0.59

 Classic

-0.59

 classics

-0.58

laughs

-0.55

 bombshell

-0.53

forward

-0.51

IBLE

-0.50

 Canary

-0.50

 levers

-0.48

LR

-0.48

POSITIVE LOGITS

most

0.73

 nowhere

0.70

few

0.68

zig

0.66

mire

0.65

nia

0.65

phans

0.64

phan

0.63

nothing

0.63

worst

0.63

Activations Density 0.525%

information related to numbers, especially rankings or statistics

phrases related to numerical data and statistics

No Comments

No Known Activations