INDEX

Explanations

proper nouns or names with common words in between

oai_token-act-pair · gpt-3.5-turbo

references to brands or entities in the context of discussions about products or consumer behavior

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ÂŃ

-0.86

 Amtrak

-0.57

earchers

-0.53

Washington

-0.53

hower

-0.52

ifestyle

-0.52

 fracking

-0.51

 Washington

-0.51

orest

-0.51

vernment

-0.51

POSITIVE LOGITS

[+

0.73

canon

0.69

 plagiar

0.64

 Variant

0.63

 english

0.61

 deleted

0.61

 deletion

0.60

 verse

0.60

 cipher

0.60

 alot

0.59

Activations Density 2.263%

proper nouns or names with common words in between

references to brands or entities in the context of discussions about products or consumer behavior

No Comments

No Known Activations