INDEX

Explanations

instances where something is suggested or stated

oai_token-act-pair · gpt-3.5-turbo

references to signs or signals suggesting evidence or implications

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

venge

-0.95

ctors

-0.86

bers

-0.86

isol

-0.81

rit

-0.77

iling

-0.76

fare

-0.76

vas

-0.74

blogs

-0.73

awar

-0.73

POSITIVE LOGITS

 indicated

1.17

 indications

1.06

 indicates

1.01

 indicate

0.91

 indication

0.89

 intervals

0.86

 indicating

0.85

 hints

0.83

 signaled

0.77

WER

0.74

Activations Density 0.009%

instances where something is suggested or stated

references to signs or signals suggesting evidence or implications

No Comments

No Known Activations