INDEX

Explanations

negative interactions and conflicts between characters

oai_token-act-pair · gpt-3.5-turbo

instances of body shaming and mockery in social interactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 contempl

-0.65

erning

-0.65

 modesty

-0.64

 suitable

-0.62

igmatic

-0.60

icip

-0.59

particularly

-0.57

 gazing

-0.57

 gaze

-0.56

igent

-0.55

POSITIVE LOGITS

 fucked

0.89

 didnt

0.88

 THEN

0.87

 gonna

0.86

 blah

0.85

 yeah

0.85

 fuckin

0.85

uh

0.84

 shit

0.81

Yeah

0.80

Activations Density 1.565%

negative interactions and conflicts between characters

instances of body shaming and mockery in social interactions

No Comments

No Known Activations