INDEX

Explanations

references to the media network "ABC"

oai_token-act-pair · gpt-3.5-turbo

mentions of the ABC network

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

amoto

-0.74

irts

-0.73

boarding

-0.71

agues

-0.67

wagen

-0.64

oppable

-0.64

ainer

-0.64

 sclerosis

-0.64

vik

-0.63

umatic

-0.63

POSITIVE LOGITS

DEF

1.46

NEWS

0.95

IS

0.85

nect

0.81

IRO

0.80

 News

0.77

elaide

0.72

ABC

0.72

RL

0.72

News

0.72

Activations Density 0.009%

references to the media network "ABC"

mentions of the ABC network

No Comments

No Known Activations