INDEX

Explanations

words related to qualities or attributes denoted by the suffix '-ness', such as "deservingness" or "happiness"

oai_token-act-pair · gpt-3.5-turbo

concepts related to the state of being or qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 dime

-0.68

reon

-0.65

zynski

-0.61

TOR

-0.60

 major

-0.60

cyl

-0.59

JECT

-0.58

ther

-0.57

kef

-0.57

olan

-0.57

POSITIVE LOGITS

ness

1.19

fulness

1.10

iness

1.10

nesses

1.04

eness

0.93

uality

0.91

ibility

0.90

lessness

0.88

ity

0.86

NESS

0.82

Activations Density 0.012%

words related to qualities or attributes denoted by the suffix '-ness', such as "deservingness" or "happiness"

concepts related to the state of being or qualities

No Comments

No Known Activations