INDEX

Explanations

entities or groups of interest mentioned in news articles

oai_token-act-pair · gpt-3.5-turbo

the word "some" and its variations, indicating a focus on quantifying uncertainty or vagueness

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ocene

-0.78

tainment

-0.70

uyomi

-0.68

gon

-0.64

meet

-0.64

ean

-0.62

jac

-0.62

hips

-0.62

 laun

-0.61

livion

-0.61

POSITIVE LOGITS

body

0.93

how

0.93

ones

0.91

 semblance

0.90

place

0.85

 unspecified

0.82

 provinces

0.73

HOW

0.71

 mornings

0.69

 unexplained

0.69

Activations Density 0.108%

entities or groups of interest mentioned in news articles

the word "some" and its variations, indicating a focus on quantifying uncertainty or vagueness

No Comments

No Known Activations