INDEX

Explanations

mentions of specific names or terms related to law enforcement and legal cases

oai_token-act-pair · gpt-3.5-turbo

instances of the name "Nar" and related terms, possibly focusing on a specific narrative or character context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Pixie

-0.89

 Democr

-0.84

IBLE

-0.79

éĹĺ

-0.78

 Canary

-0.75

ORED

-0.72

 Wink

-0.70

xual

-0.66

 VIDEOS

-0.66

 Giuliani

-0.64

POSITIVE LOGITS

ration

1.00

Nar

0.99

ayan

0.96

rils

0.95

adh

0.89

aband

0.89

abus

0.88

aya

0.88

vik

0.88

opa

0.87

Activations Density 0.005%

mentions of specific names or terms related to law enforcement and legal cases

instances of the name "Nar" and related terms, possibly focusing on a specific narrative or character context

No Comments

No Known Activations