INDEX

Explanations

examples of the word "boring" with varying degrees of intensity

oai_token-act-pair · gpt-3.5-turbo

references to boredom

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orthy

-0.77

riott

-0.77

uana

-0.75

ports

-0.74

avia

-0.72

OURCE

-0.72

iasm

-0.70

owners

-0.70

ossession

-0.68

arta

-0.66

POSITIVE LOGITS

 boring

0.94

 crap

0.88

ly

0.86

glers

0.83

 quot

0.80

 stale

0.80

 mundane

0.75

 filler

0.75

bye

0.74

 lame

0.73

Activations Density 0.013%

examples of the word "boring" with varying degrees of intensity

references to boredom

No Comments

No Known Activations