INDEX

Explanations

terms related to official government offices or departments

oai_token-act-pair · gpt-3.5-turbo

references to government offices or 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ushima

-0.72

nir

-0.70

storms

-0.69

 Haram

-0.68

hered

-0.67

ervative

-0.67

 Kaine

-0.67

uador

-0.66

udeau

-0.65

ipe

-0.65

POSITIVE LOGITS

0.97

 Depot

0.87

 Office

0.85

gur

0.72

Office

0.71

 Paper

0.68

 Desk

0.68

 Square

0.68

velop

0.67

 Procedure

0.66

Activations Density 0.017%

terms related to official government offices or departments

references to government offices or agencies

No Comments

No Known Activations