INDEX

Explanations

titles or positions of authority or leadership in organizations

oai_token-act-pair · gpt-3.5-turbo

phrases that reference leadership roles or titles within organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Required

-0.74

SPONSORED

-0.73

Synopsis

-0.72

 stub

-0.71

reci

-0.70

 blew

-0.68

 excuse

-0.66

 behaved

-0.66

 fuck

-0.65

 proved

-0.64

POSITIVE LOGITS

 advocacy

0.85

 Greenpeace

0.80

 Aval

0.76

 consultancy

0.75

 Concern

0.75

 Sierra

0.75

 Ceres

0.75

 Humane

0.75

 Confederation

0.73

 Libre

0.73

Activations Density 0.085%

titles or positions of authority or leadership in organizations

phrases that reference leadership roles or titles within organizations

No Comments

No Known Activations