INDEX

Explanations

adjectives related to characteristics or qualities

oai_token-act-pair · gpt-3.5-turbo

references to various forms of written content or narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

xtap

-0.67

 Seventh

-0.65

 sugg

-0.61

 Phill

-0.60

 Branch

-0.59

Pru

-0.58

verty

-0.57

å¼

-0.57

 Tick

-0.56

POSITIVE LOGITS

 nonetheless

1.10

 itself

0.98

 anyway

0.78

 ourselves

0.77

 anyways

0.76

 oneself

0.76

 alike

0.75

nesses

0.75

ain

0.74

 herself

0.73

Activations Density 0.414%

adjectives related to characteristics or qualities

references to various forms of written content or narratives

No Comments

No Known Activations