INDEX

Explanations

descriptive words related to senses and physical characteristics

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

elements related to physical conditions and descriptions of appearance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 Examples

-0.72

akes

-0.72

_>

-0.71

bia

-0.71

 Calls

-0.71

enthal

-0.69

roups

-0.69

Factor

-0.69

%:

-0.69

 Contracts

-0.69

POSITIVE LOGITS

 bearded

0.96

 angular

0.93

 wrink

0.92

 lifeless

0.91

 sweaty

0.90

 sandy

0.89

 dusty

0.87

 oily

0.86

 skeletal

0.84

 translucent

0.84

Activations Density 0.137%

descriptive words related to senses and physical characteristics

elements related to physical conditions and descriptions of appearance

No Comments

No Known Activations