INDEX

Explanations

words related to making something stricter or more secure

oai_token-act-pair · gpt-3.5-turbo

terms related to increasing restrictions or regulations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Glory

-0.77

 Columb

-0.72

mits

-0.71

 Lima

-0.71

agne

-0.70

 Mamm

-0.69

 Americas

-0.69

 Literary

-0.69

ULT

-0.68

Flu

-0.67

POSITIVE LOGITS

 tightening

1.48

 tightened

1.45

 tighten

1.35

 undermin

1.05

 tension

1.04

 tighter

1.03

 loosen

1.03

 easing

0.97

 clamp

0.97

 loos

0.91

Activations Density 0.007%

words related to making something stricter or more secure

terms related to increasing restrictions or regulations

No Comments

No Known Activations