INDEX

Explanations

terms related to legal and political discourse

oai_token-act-pair · gpt-3.5-turbo

arguments against generalized assumptions regarding rights and responsibilities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Billion

-0.79

 Emails

-0.77

 Lucky

-0.76

wow

-0.73

 Crazy

-0.72

 Clicker

-0.72

 Pump

-0.71

 Sold

-0.71

roller

-0.70

 PARK

-0.70

POSITIVE LOGITS

 subord

1.27

 distinctions

1.27

 normative

1.19

 exclusion

1.19

 privile

1.18

 recourse

1.15

 adherence

1.15

 conformity

1.15

 metaphysical

1.13

 reciprocal

1.12

Activations Density 0.424%

terms related to legal and political discourse

arguments against generalized assumptions regarding rights and responsibilities

No Comments

No Known Activations