INDEX

Explanations

phrases related to important information or guides

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases indicating necessity or guidance for the reader

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

aeper

-0.65

ythm

-0.64

Vu

-0.64

ween

-0.63

sie

-0.62

gue

-0.60

vertisement

-0.60

gas

-0.60

gur

-0.59

acity

-0.58

POSITIVE LOGITS

 need

1.15

 NEED

1.09

'll

1.05

need

1.03

 SHOULD

1.00

 Need

0.98

 MUST

0.94

 should

0.88

 gotta

0.88

 Should

0.84

Activations Density 0.088%

phrases related to important information or guides

phrases indicating necessity or guidance for the reader

No Comments

No Known Activations