INDEX

Explanations

phrases related to different approaches or methods

oai_token-act-pair · gpt-3.5-turbo

expressions of different methods or strategies

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

horn

-0.82

*)

-0.74

hai

-0.72

ports

-0.71

rings

-0.70

pool

-0.70

egg

-0.69

 Cosponsors

-0.68

checks

-0.67

usercontent

-0.65

POSITIVE LOGITS

 overcome

0.92

 maximize

0.87

 avoid

0.86

 achieve

0.85

 solve

0.82

 ensure

0.82

 minimize

0.81

 counteract

0.80

 accomplish

0.78

 navigate

0.78

Activations Density 0.134%

phrases related to different approaches or methods

expressions of different methods or strategies

No Comments

No Known Activations