INDEX

Explanations

locations or places with a common term 'San' followed by a variety of words

oai_token-act-pair · gpt-3.5-turbo

references to the word "San" relating to various cities or regions, particularly those in California

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

theless

-0.77

ï¸ı

-0.71

ilater

-0.69

llor

-0.68

 confir

-0.66

 clipboard

-0.66

 Wink

-0.65

llers

-0.64

 monop

-0.63

 homework

-0.63

POSITIVE LOGITS

 Francisco

1.12

San

1.11

 Diego

1.05

ction

0.99

ctuary

0.98

 Antonio

0.98

ibel

0.92

Disk

0.89

gha

0.88

itary

0.87

Activations Density 0.017%

locations or places with a common term 'San' followed by a variety of words

references to the word "San" relating to various cities or regions, particularly those in California

No Comments

No Known Activations