INDEX

Explanations

references to office spaces and settings

oai_token-act-pair · gpt-3.5-turbo

mentions of "office" or "offices" in various contexts

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

oise

-0.90

theless

-0.81

ebook

-0.74

ï¸

-0.73

âķĲâķĲ

-0.73

 Adin

-0.70

ivals

-0.69

Pause

-0.69

Prev

-0.69

eful

-0.66

POSITIVE LOGITS

tops

1.04

holder

0.97

holders

0.95

0.83

gur

0.81

 clerk

0.79

 tower

0.77

 desk

0.73

person

0.73

 holder

0.73

Activations Density 0.025%

references to office spaces and settings

mentions of "office" or "offices" in various contexts

No Comments

No Known Activations