INDEX

Explanations

titles or roles of high-ranking officials or executives

oai_token-act-pair · gpt-3.5-turbo

terms related to leadership or high-ranking positions

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

[|

-0.78

rawdownloadcloneembedreportprint

-0.72

 Caption

-0.70

pload

-0.68

 Centauri

-0.64

 Latter

-0.64

 Memories

-0.62

âĶĢâĶĢâĶĢâĶĢ

-0.61

 Waves

-0.61

 Tags

-0.60

POSITIVE LOGITS

 overseeing

0.80

Edu

0.80

 Catherine

0.74

 Sergio

0.74

 Steph

0.73

 Roger

0.72

 Stephan

0.72

 Robert

0.72

ials

0.71

 Anita

0.71

Activations Density 0.182%

titles or roles of high-ranking officials or executives

terms related to leadership or high-ranking positions

No Comments

No Known Activations