INDEX

Explanations

words related to the emotional tone or attitude expressed in text

oai_token-act-pair · gpt-3.5-turbo

references to tone in discussions or narratives

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

minster

-0.96

iott

-0.83

icipated

-0.80

inosaur

-0.78

arus

-0.75

riks

-0.75

ordan

-0.75

abis

-0.73

ernels

-0.71

essions

-0.70

POSITIVE LOGITS

 deaf

1.21

 tone

1.06

 Tone

0.95

 assurance

0.92

 tones

0.87

 tune

0.80

tone

0.73

 amplification

0.72

 sensitivity

0.71

 dial

0.69

Activations Density 0.017%

words related to the emotional tone or attitude expressed in text

references to tone in discussions or narratives

No Comments

No Known Activations