INDEX

Explanations

superlatives regarding quality or excellence

oai_token-act-pair · gpt-3.5-turbo

references to the best or highest quality in various contexts

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

idon

-0.96

zanne

-0.73

ustom

-0.73

reon

-0.72

aryn

-0.71

mitting

-0.70

 actionGroup

-0.69

ategory

-0.69

clair

-0.68

vim

-0.67

POSITIVE LOGITS

seller

1.29

iary

1.07

sell

0.91

 suited

0.90

iaries

0.89

hest

0.88

 possible

0.82

 behaved

0.81

est

0.79

 predictor

0.78

Activations Density 0.045%

superlatives regarding quality or excellence

references to the best or highest quality in various contexts

No Comments

No Known Activations