INDEX

Explanations

adverbs ending in 'ly' with a high level of activation

oai_token-act-pair · gpt-3.5-turbo

adverbs that describe the manner in which actions are performed

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Unch

-0.57

 NETWORK

-0.54

ACC

-0.53

BIP

-0.52

RECT

-0.51

 Confederation

-0.51

 fateful

-0.50

adr

-0.50

ODUCT

-0.49

 purported

-0.48

POSITIVE LOGITS

ly

4.25

LY

2.85

liness

2.07

lys

2.03

liest

1.96

lly

1.82

lier

1.74

lies

1.74

fully

1.64

ELY

1.50

Activations Density 0.043%

adverbs ending in 'ly' with a high level of activation

adverbs that describe the manner in which actions are performed

No Comments

No Known Activations