INDEX

Explanations

phrases relating to comparison or inclusion as well as some verbs

oai_token-act-pair · gpt-3.5-turbo

phrases indicating comparisons or contrasts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ãĥ

-0.43

usk

-0.43

 ÂŃ

-0.42

ï¿½ï¿½ï¿½

-0.42

Gra

-0.41

Sport

-0.40

Aren

-0.40

Euro

-0.38

ath

-0.38

egu

-0.38

POSITIVE LOGITS

 embodiments

0.57

 secondly

0.52

 thence

0.50

 optionally

0.50

 verbs

0.49

 hereafter

0.49

 ital

0.48

 endif

0.47

 specifying

0.46

 recomp

0.46

Activations Density 1.068%

phrases relating to comparison or inclusion as well as some verbs

phrases indicating comparisons or contrasts

No Comments

No Known Activations