INDEX

Explanations

newly coined terms or phrases

oai_token-act-pair · gpt-3.5-turbo

terms related to the creation or introduction of new concepts and phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

HCR

-0.69

isite

-0.66

rooms

-0.66

adesh

-0.63

alon

-0.63

phabet

-0.61

Aid

-0.61

ateral

-0.61

finder

-0.60

oyal

-0.59

POSITIVE LOGITS

 coined

1.17

ãĥ¼ãĥĨãĤ£

0.93

ãĥ¥

0.87

mint

0.83

rences

0.82

entimes

0.77

ãĥ³ãĤ¸

0.76

 invented

0.74

ãĥ£

0.73

vern

0.73

Activations Density 0.021%

newly coined terms or phrases

terms related to the creation or introduction of new concepts and phrases

No Comments

No Known Activations