INDEX

Explanations

references to high-ranking or important figures/entities within a specific context or domain

oai_token-act-pair · gpt-3.5-turbo

references to arch rivals or archbishops

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Ô

-0.84

uana

-0.82

arettes

-0.78

leneck

-0.74

 Citation

-0.70

Ã±a

-0.69

ACTION

-0.67

ABE

-0.66

OTOS

-0.65

 Serve

-0.64

POSITIVE LOGITS

bishop

1.06

ipel

1.04

 arch

0.97

di

0.87

itect

0.85

ivist

0.81

ival

0.81

rival

0.81

hum

0.79

arch

0.76

Activations Density 0.008%

references to high-ranking or important figures/entities within a specific context or domain

references to arch rivals or archbishops

No Comments

No Known Activations