INDEX

Explanations

phrases related to policies, organizations, and official actions

oai_token-act-pair · gpt-3.5-turbo

terms related to government actions and policies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

emaker

-0.72

YC

-0.72

asio

-0.69

rise

-0.69

eton

-0.68

tops

-0.68

deck

-0.67

steen

-0.67

caster

-0.66

lia

-0.66

POSITIVE LOGITS

Act

0.85

 Commission

0.81

 Institute

0.81

 Organization

0.77

 Fund

0.77

 Agreement

0.76

 Tribunal

0.76

 Framework

0.75

 Initiative

0.75

 Agency

0.75

Activations Density 0.145%

phrases related to policies, organizations, and official actions

terms related to government actions and policies

No Comments

No Known Activations