INDEX

Explanations

words related to negative or derogatory descriptions of people or actions

oai_token-act-pair · gpt-3.5-turbo

terms related to deceitful or manipulative behavior

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

 neoc

-0.75

 lapse

-0.70

 Ramadan

-0.66

 craving

-0.66

 famine

-0.65

theless

-0.64

 millennium

-0.64

 foremost

-0.63

 abundantly

-0.62

 inaug

-0.62

POSITIVE LOGITS

cheon

0.95

hett

0.81

udo

0.79

berus

0.77

itzer

0.76

anut

0.75

uli

0.75

anon

0.74

oslav

0.71

arios

0.70

Activations Density 0.079%

words related to negative or derogatory descriptions of people or actions

terms related to deceitful or manipulative behavior

No Comments

No Known Activations