INDEX

Explanations

phrases denoting exceptions

oai_token-act-pair · gpt-3.5-turbo

words or phrases signaling exceptions

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

entry

-0.63

natureconservancy

-0.63

mary

-0.62

acia

-0.61

 Venture

-0.60

 creep

-0.59

ve

-0.58

ushed

-0.58

ooming

-0.57

uers

-0.57

POSITIVE LOGITS

soever

1.04

 except

0.97

withstanding

0.88

except

0.84

 insofar

0.79

heses

0.78

 incidentally

0.77

ional

0.74

unless

0.73

 namely

0.70

Activations Density 0.011%

phrases denoting exceptions

words or phrases signaling exceptions

No Comments

No Known Activations