INDEX

Explanations

words related to criticism or negative judgment

oai_token-act-pair · gpt-3.5-turbo

terms that convey negative judgments or criticisms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vals

-0.71

chance

-0.68

quart

-0.65

semble

-0.65

rollers

-0.63

interrupted

-0.62

eon

-0.61

ngth

-0.61

 Cups

-0.60

asio

-0.60

POSITIVE LOGITS

 underest

0.88

 enough

0.85

 underestimate

0.79

 overest

0.74

 hypocr

0.73

 because

0.72

uably

0.71

 grounds

0.71

 hypocrisy

0.70

 exagger

0.70

Activations Density 0.165%

words related to criticism or negative judgment

terms that convey negative judgments or criticisms

No Comments

No Known Activations