INDEX

Explanations

words related to experiences or abstract concepts

oai_token-act-pair · gpt-3.5-turbo

variations of the word "experience"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 approved

-0.63

 charged

-0.61

 penetrating

-0.60

 eligible

-0.60

 legally

-0.60

 authorized

-0.59

OK

-0.59

 unilaterally

-0.59

 revision

-0.58

 initiated

-0.58

POSITIVE LOGITS

ience

4.78

iences

3.17

iency

2.63

ienced

2.21

IENCE

2.21

ient

2.21

ients

1.83

iance

1.73

iencies

1.64

iment

1.50

Activations Density 0.009%

words related to experiences or abstract concepts

variations of the word "experience"

No Comments

No Known Activations