INDEX

Explanations

phrases related to government offices or departments and managing officials

oai_token-act-pair · gpt-3.5-turbo

references to government offices and agencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ashtra

-1.05

oway

-0.84

iland

-0.76

agher

-0.74

otaur

-0.72

ulia

-0.71

plings

-0.70

igue

-0.70

odka

-0.70

iflower

-0.69

POSITIVE LOGITS

 Strategic

0.76

 Personnel

0.76

 Duty

0.74

Correction

0.73

 Inquiry

0.73

 Vice

0.72

 Directors

0.70

ãĤ±

0.70

 Budget

0.69

 Governors

0.68

Activations Density 0.133%

phrases related to government offices or departments and managing officials

references to government offices and agencies

No Comments

No Known Activations