INDEX

Explanations

phrases related to feedback and instructions

oai_token-act-pair · gpt-3.5-turbo

phrases related to opinions and feedback

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

cffffcc

-0.76

 overtake

-0.67

 scares

-0.67

 emerges

-0.66

 bothers

-0.66

 shocks

-0.66

luaj

-0.66

 billions

-0.65

 stake

-0.64

POSITIVE LOGITS

 hereby

1.12

 available

0.91

 provided

0.88

TBD

0.88

 appreciated

0.88

 advised

0.85

 tentative

0.84

 LIMITED

0.83

 reviewed

0.82

 emailed

0.81

Activations Density 0.256%

phrases related to feedback and instructions

phrases related to opinions and feedback

No Comments

No Known Activations