INDEX

Explanations

phrases related to addressing a public audience

oai_token-act-pair · gpt-3.5-turbo

expressions of frustration or emphasis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lishes

-0.65

Prev

-0.65

-0.63

 plagued

-0.62

 varied

-0.59

arcity

-0.58

COMPLE

-0.58

 influences

-0.58

accessible

-0.57

olves

-0.57

POSITIVE LOGITS

 congr

1.09

 congratulations

1.01

ratulations

1.00

 gentlemen

1.00

bye

0.95

 welcome

0.91

give

0.89

 please

0.89

sir

0.88

huh

0.88

Activations Density 0.804%

phrases related to addressing a public audience

expressions of frustration or emphasis

No Comments

No Known Activations