INDEX

Explanations

written documents such as letters, memos, articles, or blog posts

oai_token-act-pair · gpt-3.5-turbo

references to various types of documents and communications, such as letters and memos

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

instead

-0.67

cause

-0.61

 outweigh

-0.60

ctrl

-0.60

artifacts

-0.58

 injust

-0.58

tics

-0.56

illard

-0.56

 despise

-0.56

animate

-0.56

POSITIVE LOGITS

 nutshell

0.84

idav

0.74

 announcing

0.73

 interview

0.69

 titled

0.68

 HuffPost

0.67

 emailed

0.65

fter

0.64

 published

0.64

 released

0.64

Activations Density 0.120%

written documents such as letters, memos, articles, or blog posts

references to various types of documents and communications, such as letters and memos

No Comments

No Known Activations