INDEX

Explanations

phrases related to physical actions and objects

oai_token-act-pair · gpt-3.5-turbo

indefinite articles and descriptors for various objects or actions

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

tests

-0.82

Events

-0.78

Sources

-0.78

grounds

-0.77

ablishment

-0.76

states

-0.74

Init

-0.73

align

-0.73

iments

-0.72

 relies

-0.72

POSITIVE LOGITS

 handful

1.21

 bunch

1.16

 couple

1.01

few

1.00

 dozen

1.00

 pair

0.99

 flurry

0.98

 small

0.96

 towel

0.93

 knife

0.92

Activations Density 0.464%

phrases related to physical actions and objects

indefinite articles and descriptors for various objects or actions

No Comments

No Known Activations