INDEX

Explanations

phrases and questions related to thoughts or considerations

oai_token-act-pair · gpt-3.5-turbo

references to personal experiences and inquiries about actions

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

 Brush

-0.66

tops

-0.66

ãĥĥ

-0.62

oubt

-0.62

amy

-0.61

 Bliss

-0.61

itude

-0.61

AIR

-0.60

ACY

-0.59

 Frazier

-0.59

POSITIVE LOGITS

ribed

0.82

 fared

0.78

 unfolded

0.77

 structured

0.77

 behave

0.75

 interpreted

0.74

 stacked

0.73

 manifests

0.70

 behaved

0.69

 differs

0.68

Activations Density 0.177%

phrases and questions related to thoughts or considerations

references to personal experiences and inquiries about actions

No Comments

No Known Activations