INDEX

Explanations

phrases related to sequences or orders

oai_token-act-pair · gpt-3.5-turbo

references to chronological or sequential ordering

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

tek

-1.01

mouth

-0.79

erville

-0.74

erred

-0.72

 Valkyrie

-0.69

ouf

-0.67

WT

-0.66

unc

-0.65

erm

-0.64

ako

-0.64

POSITIVE LOGITS

 chronological

1.37

 alphabet

1.21

 ascending

1.20

 descending

1.06

 comma

0.99

 sorting

0.93

 ordering

0.92

 consecut

0.89

 chron

0.88

 alternating

0.88

Activations Density 0.346%

phrases related to sequences or orders

references to chronological or sequential ordering

No Comments

No Known Activations