INDEX

Explanations

words related to inciting or triggering reactions or responses, especially in a negative context

oai_token-act-pair · gpt-3.5-turbo

phrases that provoke a reaction or discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 afford

-0.70

 undone

-0.69

 Paddock

-0.65

 Anat

-0.64

 survived

-0.61

Aj

-0.61

 OnePlus

-0.60

 disapp

-0.59

ACTED

-0.58

matched

-0.58

POSITIVE LOGITS

ire

0.90

 wrath

0.89

inki

0.84

 curiosity

0.84

 creativity

0.80

 formation

0.78

 growth

0.76

 frenzy

0.74

 skepticism

0.72

 enthusi

0.71

Activations Density 0.314%

words related to inciting or triggering reactions or responses, especially in a negative context

phrases that provoke a reaction or discussion

No Comments

No Known Activations