INDEX

Explanations

phrases related to logic, reasoning, and common sense

oai_token-act-pair · gpt-3.5-turbo

references to "common sense" and related concepts

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

letes

-0.78

zona

-0.70

 Become

-0.68

otos

-0.65

asus

-0.65

dyl

-0.65

=-=-

-0.65

orthy

-0.63

wards

-0.62

atoon

-0.62

POSITIVE LOGITS

wealth

1.64

alities

1.39

ality

1.28

 denomin

1.21

places

1.15

 sense

1.11

 decency

1.07

 ancestor

0.98

place

0.98

alty

0.93

Activations Density 0.047%

phrases related to logic, reasoning, and common sense

references to "common sense" and related concepts

No Comments

No Known Activations