INDEX

Explanations

mentions of podcasts and podcast hosts

oai_token-act-pair · gpt-3.5-turbo

references to podcasts and their hosts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 ÂŃ

-0.82

ÂŃ

-0.65

 elector

-0.65

 diplomats

-0.64

 detainees

-0.64

xes

-0.64

 stricken

-0.64

 statute

-0.63

 compromises

-0.62

 expectancy

-0.61

POSITIVE LOGITS

 Youtube

0.97

 Patreon

0.96

Awesome

0.91

 Awesome

0.90

 youtube

0.88

:)

0.86

;)

0.86

outube

0.81

 facebook

0.77

 http

0.77

Activations Density 0.850%

mentions of podcasts and podcast hosts

references to podcasts and their hosts

No Comments

No Known Activations