INDEX

Explanations

phrases related to warnings or caution

oai_token-act-pair · gpt-3.5-turbo

terms related to warnings and alerts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

inion

-0.72

anova

-0.70

earch

-0.68

iop

-0.66

ourced

-0.66

 Legends

-0.66

ttle

-0.65

pora

-0.64

á

-0.64

morph

-0.62

POSITIVE LOGITS

 signs

1.22

 Signs

1.20

 labels

1.07

 bells

1.04

 warnings

0.93

 letters

0.91

 signals

0.91

 notices

0.90

 label

0.87

 sign

0.85

Activations Density 0.057%

phrases related to warnings or caution

terms related to warnings and alerts

No Comments

No Known Activations