INDEX

Explanations

adjectives related to ranking and evaluation

oai_token-act-pair · gpt-3.5-turbo

words and phrases that indicate ratings or evaluations

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

ODUCT

-0.74

 clar

-0.69

 Extract

-0.66

 extracts

-0.66

 else

-0.65

 acquaint

-0.62

 evict

-0.62

 extracting

-0.61

 Clar

-0.61

 extract

-0.60

POSITIVE LOGITS

rated

3.53

rations

1.94

ration

1.58

rator

1.51

 rated

1.34

graded

1.25

rators

1.25

Rated

1.25

valued

1.20

played

1.17

Activations Density 0.010%

adjectives related to ranking and evaluation

words and phrases that indicate ratings or evaluations

No Comments

No Known Activations