INDEX

Explanations

contractions with 'd' like "I'd"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the contraction "I'd"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

aders

-0.68

PI

-0.67

IDA

-0.65

å£«

-0.63

 Gutenberg

-0.63

strings

-0.60

åŃ

-0.60

 Telecommunications

-0.60

 Hels

-0.60

æ°

-0.59

POSITIVE LOGITS

 been

1.05

 gotten

1.01

 gotta

1.01

 gone

0.92

been

0.83

 gladly

0.79

 forgotten

0.77

hran

0.77

 Been

0.77

terness

0.77

Activations Density 0.017%

contractions with 'd' like "I'd"

occurrences of the contraction "I'd"

No Comments

No Known Activations