INDEX

Explanations

exclamatory phrases related to actions or items

oai_token-act-pair · gpt-3.5-turbo

punctuations or exclamatory and interrogative expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

rab

-0.67

sonian

-0.66

oria

-0.66

ral

-0.65

zbollah

-0.65

matically

-0.63

eling

-0.62

bern

-0.62

romy

-0.61

gging

-0.61

POSITIVE LOGITS

âĶģ

0.77

 srfAttach

0.75

uits

0.73

ategory

0.69

#$

0.67

uly

0.66

 Anyway

0.66

theless

0.66

ittens

0.63

ugg

0.63

Activations Density 0.017%

exclamatory phrases related to actions or items

punctuations or exclamatory and interrogative expressions

No Comments

No Known Activations