INDEX

Explanations

references to various generations (Gen) followed by a number

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Gen," likely referring to generational topics or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

 spirited

-0.62

 form

-0.62

 hitch

-0.61

 seasonal

-0.59

 airborne

-0.58

 dogged

-0.57

 Journey

-0.56

 divided

-0.56

 hours

-0.55

selves

-0.55

POSITIVE LOGITS

uine

1.56

iuses

1.45

ocide

1.32

etically

1.31

esis

1.26

omics

1.17

etics

1.15

zyme

1.13

otype

1.09

etic

1.08

Activations Density 0.036%

references to various generations (Gen) followed by a number

mentions of the term "Gen," likely referring to generational topics or issues

No Comments

No Known Activations