INDEX

Explanations

specific phrases related to the application or relevance of information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating applicability or relevance to various contexts

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

ikk

-0.74

GOP

-0.72

erg

-0.69

inus

-0.66

manac

-0.65

ocative

-0.64

ipedia

-0.64

trap

-0.64

inder

-0.64

ãĥīãĥ©

-0.63

POSITIVE LOGITS

 EVERY

1.20

 every

1.18

 both

1.18

 virtually

1.14

 everyone

1.12

 everything

1.11

all

1.11

 everybody

1.10

 only

1.08

ALL

1.06

Activations Density 0.533%

specific phrases related to the application or relevance of information

phrases indicating applicability or relevance to various contexts

No Comments

No Known Activations