INDEX

Explanations

phrases related to taking actions or steps

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "the" indicating a focus on articles

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-+-+

-0.82

lished

-0.81

tions

-0.79

alde

-0.77

tion

-0.77

lich

-0.76

Operation

-0.76

ambo

-0.74

cade

-0.73

ntil

-0.73

POSITIVE LOGITS

 brunt

1.35

 plunge

1.34

 opportunity

1.16

 reins

1.15

 initiative

1.12

 helm

1.10

 bait

1.07

 blame

1.07

 liberty

1.04

cue

0.95

Activations Density 0.049%

phrases related to taking actions or steps

repeated mentions of the word "the" indicating a focus on articles

No Comments

No Known Activations