INDEX

Explanations

keywords related to specific actions or events

oai_token-act-pair · gpt-3.5-turbo

references to action or activism

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gling

-0.67

mbuds

-0.65

conservancy

-0.64

YL

-0.63

gins

-0.63

DAQ

-0.62

yth

-0.62

artifacts

-0.61

 abundant

-0.59

 quarters

-0.59

POSITIVE LOGITS

ivated

0.91

tec

0.83

 Replay

0.81

ual

0.81

ality

0.79

ulatory

0.79

ivation

0.78

aries

0.77

able

0.77

aires

0.76

Activations Density 0.037%

keywords related to specific actions or events

references to action or activism

No Comments

No Known Activations