INDEX

Explanations

phrases related to legal or disciplinary actions

oai_token-act-pair · gpt-3.5-turbo

terms related to punishment and penalties

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ITNESS

-0.77

ertility

-0.73

omorphic

-0.73

omorph

-0.72

 livest

-0.69

utherland

-0.66

atters

-0.65

secure

-0.65

pora

-0.65

clus

-0.64

POSITIVE LOGITS

 levied

1.30

 imposed

1.12

 penalties

1.06

 sanction

0.99

alties

0.96

 punishments

0.95

 penalty

0.91

 inflicted

0.90

 punishment

0.87

 leveled

0.85

Activations Density 0.115%

phrases related to legal or disciplinary actions

terms related to punishment and penalties

No Comments

No Known Activations