INDEX

Explanations

links or mentions of the company Google

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Google."

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

mble

-0.96

rone

-0.77

quez

-0.74

ivari

-0.73

sole

-0.71

philis

-0.69

alid

-0.69

stood

-0.68

iotic

-0.68

rones

-0.67

POSITIVE LOGITS

plex

0.95

 Scholar

0.91

 Maps

0.89

 Doodle

0.83

Maps

0.81

geist

0.75

 Chrome

0.75

 google

0.72

usercontent

0.72

 Google

0.71

Activations Density 0.015%

links or mentions of the company Google

mentions of the term "Google."

No Comments

No Known Activations