INDEX

Explanations

phrases related to restrictions or rules

oai_token-act-pair · gpt-3.5-turbo

negations and phrases indicating restrictions or limitations

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

Dear

-0.76

TIME

-0.73

Kings

-0.70

LOS

-0.70

WAY

-0.70

 Conversation

-0.70

 realities

-0.66

HAHA

-0.66

 Shades

-0.64

facts

-0.64

POSITIVE LOGITS

 necessarily

1.30

icable

0.87

 recommended

0.87

 removable

0.85

 overly

0.79

epad

0.77

 mandatory

0.77

 deprecated

0.76

 adjustable

0.76

icably

0.75

Activations Density 0.338%

phrases related to restrictions or rules

negations and phrases indicating restrictions or limitations

No Comments

No Known Activations