INDEX

Explanations

references to LinkedIn

oai_token-act-pair · gpt-3.5-turbo

mentions of specific platforms and media outlets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Rudolph

-0.78

fu

-0.69

 Valentine

-0.69

 Manson

-0.65

 innocuous

-0.64

 eater

-0.62

 sinister

-0.61

rav

-0.60

 Adolf

-0.60

nas

-0.59

POSITIVE LOGITS

Desk

4.04

 Desk

2.13

entimes

1.43

 LinkedIn

1.43

Morning

1.29

Monitor

1.08

mbudsman

1.06

 POLITICO

1.00

 desks

0.96

 Journey

0.90

Activations Density 0.038%

references to LinkedIn

mentions of specific platforms and media outlets

No Comments

No Known Activations