INDEX

Explanations

phrases indicating completion or progression of a task or action

oai_token-act-pair · gpt-3.5-turbo

punctuation and special characters in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wagen

-0.76

 secretaries

-0.73

 flashing

-0.73

 imperson

-0.71

 tricked

-0.68

VIDIA

-0.68

ModLoader

-0.67

 wedd

-0.67

ensibly

-0.66

 permitting

-0.65

POSITIVE LOGITS

 Additionally

0.82

 Else

0.80

<|endoftext|>

0.79

 ----------------------------------------------------------------

0.78

 Avalanche

0.77

 Anchorage

0.77

 However

0.76

 Anyway

0.75

 Meanwhile

0.74

 UNCLASSIFIED

0.74

Activations Density 0.900%

phrases indicating completion or progression of a task or action

punctuation and special characters in the text

No Comments

No Known Activations