INDEX

Explanations

phrases related to positive affirmations or approvals

oai_token-act-pair · gpt-3.5-turbo

expressions of emotional sentiment or concern

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 undet

-0.76

 synerg

-0.71

 cone

-0.71

 cannibal

-0.69

 fermented

-0.68

occ

-0.68

 oppressed

-0.68

 engineering

-0.67

 endeav

-0.67

 equival

-0.67

POSITIVE LOGITS

Asked

1.40

Newsletter

1.38

Advertisement

1.34

Another

1.33

Indeed

1.32

Refer

1.30

Contribut

1.29

Others

1.27

Meanwhile

1.25

Topics

1.24

Activations Density 0.228%

phrases related to positive affirmations or approvals

expressions of emotional sentiment or concern

No Comments

No Known Activations