INDEX

Explanations

action verbs indicating conflict or disagreement

oai_token-act-pair · gpt-3.5-turbo

active verbs related to actions taken by officials or authorities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ãĥĺ

-0.78

washer

-0.73

Interstitial

-0.70

RIP

-0.69

 presidency

-0.67

 belongs

-0.67

die

-0.67

å¤©

-0.66

fuck

-0.66

Advertisements

-0.65

POSITIVE LOGITS

 themselves

0.88

 alike

0.85

 caution

0.81

 cite

0.78

 unanimously

0.76

 warn

0.74

bably

0.73

 anticipate

0.73

 emphasize

0.71

 envision

0.71

Activations Density 0.570%

action verbs indicating conflict or disagreement

active verbs related to actions taken by officials or authorities

No Comments

No Known Activations