INDEX

Explanations

words related to disapproval or criticism

oai_token-act-pair · gpt-3.5-turbo

words related to negative judgment or condemnation

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

emade

-0.76

ahime

-0.75

unin

-0.75

ãĥ¯

-0.74

iasm

-0.74

glers

-0.73

inn

-0.71

ãĥ³ãĤ¸

-0.71

heed

-0.71

Ã£o

-0.70

POSITIVE LOGITS

 Sparkle

0.65

Parents

0.64

 payable

0.63

osponsors

0.62

INGTON

0.62

ately

0.61

vier

0.60

 Rising

0.60

 reven

0.60

ORED

0.60

Activations Density 0.042%

words related to disapproval or criticism

words related to negative judgment or condemnation

No Known Activations