INDEX

Explanations

phrases related to monetary rewards or penalties

oai_token-act-pair · gpt-3.5-turbo

phrases indicating limits or thresholds

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Maker

-0.64

 Mate

-0.62

 Munich

-0.60

 Gadget

-0.56

Â¯Â¯Â¯Â¯

-0.56

 Marcos

-0.55

Ware

-0.55

å£«

-0.55

Cu

-0.54

 jewel

-0.54

POSITIVE LOGITS

votes

1.02

graded

0.98

grades

0.93

roots

0.92

raised

0.90

rights

0.90

grading

0.85

np

0.85

regulation

0.83

dates

0.79

Activations Density 0.057%

phrases related to monetary rewards or penalties

phrases indicating limits or thresholds

No Comments

No Known Activations