INDEX

Explanations

phrases related to the position or physical location of objects or elements

oai_token-act-pair · gpt-3.5-turbo

references to spatial arrangements and side-related distinctions

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

untarily

-0.59

NBA

-0.57

 Pengu

-0.57

ivated

-0.54

 spoiled

-0.54

etheus

-0.53

kinson

-0.53

ogo

-0.53

ufact

-0.52

 acquaintance

-0.52

POSITIVE LOGITS

stretched

0.73

 DISTRICT

0.69

akra

0.67

occupied

0.63

iHUD

0.62

Thumbnail

0.61

 territory

0.58

 edges

0.57

 Weston

0.57

 Denis

0.57

Activations Density 1.354%

phrases related to the position or physical location of objects or elements

references to spatial arrangements and side-related distinctions

No Comments

No Known Activations