INDEX

Explanations

phrases related to silencing or stopping something

oai_token-act-pair · gpt-3.5-turbo

phrases related to silencing or shutting down

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

APH

-0.72

 likeness

-0.69

pain

-0.69

 sophistication

-0.68

KNOWN

-0.67

apego

-0.66

nard

-0.66

ammy

-0.64

 maturity

-0.64

VL

-0.64

POSITIVE LOGITS

byss

0.76

 Deng

0.71

road

0.64

sec

0.61

anded

0.61

Repeat

0.61

Ped

0.61

ipedia

0.60

baugh

0.59

eye

0.59

Activations Density 0.096%

phrases related to silencing or stopping something

phrases related to silencing or shutting down

No Comments

No Known Activations