INDEX

Explanations

commands or statements urging action or change

oai_token-act-pair · gpt-3.5-turbo

commands or calls to action, particularly urging to "stop" various behaviors or practices

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

Sov

-0.87

ammy

-0.83

ighth

-0.76

aths

-0.72

ety

-0.69

ocry

-0.69

umbered

-0.67

axter

-0.67

olesc

-0.67

uth

-0.66

POSITIVE LOGITS

watching

0.96

gap

0.93

watch

0.92

 bothering

0.87

 smoking

0.81

ŃĶ

0.78

 blinking

0.77

 breathing

0.77

smoking

0.72

loading

0.72

Activations Density 0.027%

commands or statements urging action or change

commands or calls to action, particularly urging to "stop" various behaviors or practices

No Comments

No Known Activations