INDEX

Explanations

phrases related to actions causing a response or reaction

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a cause-and-effect relationship or actions that trigger responses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

mate

-0.75

atum

-0.71

sm

-0.66

rend

-0.65

lat

-0.65

ighth

-0.64

 quart

-0.63

breaker

-0.61

 route

-0.61

uded

-0.61

POSITIVE LOGITS

 laughter

0.92

 prompt

0.88

 outcry

0.88

 inquiries

0.86

 applause

0.82

 warnings

0.82

 speculation

0.80

 comparisons

0.79

 widespread

0.77

 investigations

0.76

Activations Density 0.077%

phrases related to actions causing a response or reaction

phrases indicating a cause-and-effect relationship or actions that trigger responses

No Comments

No Known Activations