INDEX

Explanations

programming-related keywords and functions

oai_token-act-pair · gpt-3.5-turbo

code-related terminology and constructs

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

 millenn

-0.82

 condem

-0.67

illac

-0.66

 Summers

-0.61

 deregulation

-0.59

 destro

-0.58

psey

-0.58

elaide

-0.57

 disadvant

-0.56

Coach

-0.56

POSITIVE LOGITS

 sqor

0.83

 suffix

0.76

ibling

0.69

onym

0.67

ymes

0.67

 Observer

0.66

 denotes

0.65

lihood

0.65

yes

0.64

cent

0.64

Activations Density 0.669%

programming-related keywords and functions

code-related terminology and constructs

No Comments

No Known Activations