INDEX

Explanations

names of individuals followed by descriptions of them or their roles

oai_token-act-pair · gpt-3.5-turbo

names of individuals associated with their roles or titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

eeds

-0.77

olk

-0.71

lings

-0.71

oor

-0.67

fw

-0.64

estone

-0.64

ilion

-0.64

alyses

-0.63

amn

-0.63

oÄŁ

-0.63

POSITIVE LOGITS

 Jeffrey

0.91

 Denis

0.90

 Coun

0.89

 Mike

0.89

 Miche

0.88

 Stephanie

0.86

 Sergio

0.86

Jim

0.85

 Gina

0.85

 Steve

0.84

Activations Density 0.117%

names of individuals followed by descriptions of them or their roles

names of individuals associated with their roles or titles

No Comments

No Known Activations