INDEX

Explanations

references to time periods, such as numbers followed by "years."

oai_token-act-pair · gpt-3.5-turbo

numerical references to time frames and durations

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

ioch

-0.69

EStream

-0.67

 speculation

-0.67

 Horizon

-0.65

 Soph

-0.64

 Machina

-0.64

 scissors

-0.64

 Valhalla

-0.62

 unfocusedRange

-0.59

 trouble

-0.59

POSITIVE LOGITS

%-

0.88

%.

0.87

teenth

0.86

%"

0.86

%:

0.86

%,

0.80

Â¢

0.77

%;

0.77

th

0.76

%]

0.76

Activations Density 0.183%

references to time periods, such as numbers followed by "years."

numerical references to time frames and durations

No Comments

No Known Activations