INDEX

Explanations

text related to physical placement or positioning of objects or body parts

oai_token-act-pair · gpt-3.5-turbo

references to physical placement or positioning 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

 Yard

-0.63

mean

-0.60

idays

-0.58

 repeated

-0.58

arson

-0.58

 folklore

-0.56

 Dhabi

-0.54

++)

-0.53

 Khalid

-0.53

 Borough

-0.52

POSITIVE LOGITS

 onto

0.87

onds

0.79

raint

0.76

iage

0.74

 cursor

0.73

ogens

0.72

pload

0.71

auts

0.69

rones

0.69

TPS

0.67

Activations Density 0.312%

text related to physical placement or positioning of objects or body parts

references to physical placement or positioning in various contexts

No Comments

No Known Activations