INDEX

Explanations

words related to personal names or proper nouns, especially those with strong emotional connotations

oai_token-act-pair · gpt-3.5-turbo

references to social and cultural themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Pwr

-0.63

 Moff

-0.59

AE

-0.55

Osw

-0.53

 Gadget

-0.50

 Colony

-0.47

 Yorkshire

-0.46

 Formula

-0.46

Gat

-0.45

Zur

-0.44

POSITIVE LOGITS

ioned

0.71

arant

0.68

odox

0.65

alin

0.62

ript

0.59

omas

0.59

omal

0.59

sing

0.59

ographed

0.58

roid

0.58

Activations Density 2.027%

words related to personal names or proper nouns, especially those with strong emotional connotations

references to social and cultural themes

No Comments

No Known Activations