INDEX

Explanations

verbs related to negative or confrontational behavior

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "engage in" followed by various activities or behaviors

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

 emerges

-0.75

 emerge

-0.65

PATH

-0.61

 shine

-0.61

makers

-0.60

loo

-0.59

 sisters

-0.59

 headlined

-0.59

 convol

-0.58

 thresholds

-0.58

POSITIVE LOGITS

ordinate

1.02

umbn

0.96

 conversation

0.89

ventory

0.82

 dialogue

0.79

 vitro

0.78

 pursuit

0.76

 discussions

0.74

 conjunction

0.73

cheon

0.73

Activations Density 0.051%

verbs related to negative or confrontational behavior

instances of the phrase "engage in" followed by various activities or behaviors

No Comments

No Known Activations