INDEX

Explanations

phrases indicating acceptance or refusal of various situations or conditions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "accept" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Peaks

-0.69

 Khalid

-0.68

oil

-0.62

storm

-0.62

waters

-0.61

jan

-0.61

Downloadha

-0.61

hur

-0.61

Bas

-0.60

 Battery

-0.60

POSITIVE LOGITS

 accept

3.38

 accepts

2.45

accept

2.33

 accepting

2.23

 accepted

2.07

Accept

2.00

 acceptance

1.97

 Accept

1.95

 reject

1.62

 tolerate

1.57

Activations Density 0.013%

phrases indicating acceptance or refusal of various situations or conditions

instances of the word "accept" in various contexts

No Comments

No Known Activations