INDEX

Explanations

names mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "named" followed by a proper noun

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

college

-0.74

hibition

-0.70

hibit

-0.65

Constructed

-0.63

atever

-0.62

 provinces

-0.62

urity

-0.62

 disparate

-0.61

 reconcile

-0.61

âķĲ

-0.60

POSITIVE LOGITS

 Desire

0.94

 Frankie

0.89

Sue

0.89

 Karl

0.88

Rex

0.88

 Theodore

0.87

 Caleb

0.86

 Harold

0.85

 Abby

0.84

 Wong

0.84

Activations Density 0.049%

names mentioned in the text

occurrences of the word "named" followed by a proper noun

No Comments

No Known Activations