INDEX

Explanations

phrases related to locations or specific places

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

occurrences of the word "se."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

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_post

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

 crooked

-0.63

 Track

-0.62

 liking

-0.61

 apprehension

-0.60

 Kirin

-0.59

 vacuum

-0.58

 incomplete

-0.57

 steep

-0.56

 mock

-0.56

 scare

-0.56

POSITIVE LOGITS

vier

1.07

gur

1.03

kai

1.01

mination

0.93

ld

0.92

hra

0.91

ktop

0.87

ngth

0.86

vo

0.86

grain

0.86

Activations Density 0.042%

phrases related to locations or specific places

occurrences of the word "se."

No Comments

No Known Activations