INDEX

Explanations

ways things are presented or formatted

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate types or methods

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

ivas

-0.81

isf

-0.78

owe

-0.77

ask

-0.76

Bridge

-0.71

NRS

-0.70

.–

-0.69

falls

-0.68

CHA

-0.68

ilver

-0.67

POSITIVE LOGITS

 Finder

0.76

 Pope

0.67

Pope

0.67

 Foss

0.65

 Prayer

0.65

 Suarez

0.65

 aiding

0.64

 Rudolph

0.64

 Hurricane

0.63

 facilitating

0.63

Activations Density 0.089%

ways things are presented or formatted

phrases that indicate types or methods

No Comments

No Known Activations