INDEX

Explanations

dates and numbers

oai_token-act-pair · gpt-3.5-turbo

elements related to dates and time-sensitive information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 stru

-0.64

Py

-0.64

UTC

-0.62

 guiActiveUnfocused

-0.62

////

-0.61

azo

-0.58

fuck

-0.58

 Tart

-0.57

ãĥĥãĥĪ

-0.57

usercontent

-0.56

POSITIVE LOGITS

 caregivers

0.81

iors

0.75

 newcomers

0.67

their

0.67

ooked

0.66

 their

0.64

 frontline

0.63

urers

0.63

 emergencies

0.62

 pets

0.61

Activations Density 1.587%

dates and numbers

elements related to dates and time-sensitive information

No Comments

No Known Activations