INDEX

Explanations

words related to injuries or wounds

oai_token-act-pair · gpt-3.5-turbo

references to sound or noise

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

isin

-0.70

å¥³

-0.61

icago

-0.61

istry

-0.60

itent

-0.60

perature

-0.59

ski

-0.59

 capital

-0.57

ISM

-0.57

 burg

-0.57

POSITIVE LOGITS

OUND

1.25

ounds

1.18

ound

0.92

Sound

0.86

Sounds

0.82

ertain

0.80

Audio

0.80

ounding

0.80

ounded

0.75

animous

0.75

Activations Density 0.010%

words related to injuries or wounds

references to sound or noise

No Comments

No Known Activations