INDEX

Explanations

locations or directions such as buildings, countries, schools, or specific places

oai_token-act-pair · gpt-3.5-turbo

locations and contexts related to escape or departure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Interstitial

-0.82

Magikarp

-0.62

Grand

-0.61

Els

-0.61

Redditor

-0.60

kw

-0.58

utm

-0.56

 specials

-0.56

Compare

-0.56

Strike

-0.56

POSITIVE LOGITS

 altogether

1.27

 prematurely

1.15

 abruptly

1.06

 entirely

0.97

 confines

0.92

 shack

0.88

 voluntarily

0.86

 onto

0.84

 unsc

0.84

 alive

0.82

Activations Density 0.324%

locations or directions such as buildings, countries, schools, or specific places

locations and contexts related to escape or departure

No Comments

No Known Activations