INDEX

Explanations

comparisons or relative quantities

oai_token-act-pair · gpt-3.5-turbo

comparative language and references to quantifiable differences

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

 Nare

-0.68

ordinate

-0.66

(?,

-0.64

enes

-0.62

 sclerosis

-0.61

":-

-0.60

ourse

-0.59

ciplinary

-0.58

 Scale

-0.57

hess

-0.57

POSITIVE LOGITS

arently

0.82

)</

0.76

-)

0.73

ngth

0.73

ãĥ¯

0.71

ardless

0.67

ãĥ¼ãĥĨãĤ£

0.66

ãĥĬ

0.62

ãĥĭ

0.62

acting

0.61

Activations Density 0.590%

comparisons or relative quantities

comparative language and references to quantifiable differences

No Comments

No Known Activations