INDEX

Explanations

numerical values in a specific range within the text

oai_token-act-pair · gpt-3.5-turbo

phrases indicating ordinal positions or instances within a sequential list

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

WER

-0.79

ertodd

-0.69

helle

-0.64

irection

-0.64

 surpr

-0.63

 condem

-0.60

heit

-0.59

overe

-0.57

cape

-0.57

 disposed

-0.56

POSITIVE LOGITS

arching

0.62

icial

0.60

frames

0.60

tan

0.60

third

0.57

their

0.57

 course

0.56

major

0.56

fourth

0.56

 thirds

0.55

Activations Density 0.103%

numerical values in a specific range within the text

phrases indicating ordinal positions or instances within a sequential list

No Comments

No Known Activations