INDEX

Explanations

content generated or created by specific news organizations

oai_token-act-pair · gpt-3.5-turbo

phrases related to news source attribution and content creation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

DeL

-0.62

escription

-0.61

cause

-0.59

nown

-0.58

essee

-0.57

iology

-0.57

uality

-0.56

 memor

-0.56

ppo

-0.56

 Examination

-0.55

POSITIVE LOGITS

 Sketch

0.75

 Pastebin

0.69

acy

0.68

ACY

0.67

 cookies

0.66

 strives

0.63

Asset

0.63

 contributors

0.61

avascript

0.60

roy

0.60

Activations Density 0.059%

content generated or created by specific news organizations

phrases related to news source attribution and content creation

No Comments

No Known Activations