INDEX

Explanations

user ratings followed by numbers

oai_token-act-pair · gpt-3.5-turbo

elements related to ratings and reviews

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

hips

-0.80

rift

-0.76

rams

-0.74

ourt

-0.69

schild

-0.69

chery

-0.67

 padd

-0.66

raint

-0.65

ruck

-0.64

hop

-0.63

POSITIVE LOGITS

Reviewer

1.43

Rated

1.08

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.91

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

0.82

 ACTIONS

0.80

ħĭ

0.77

rawdownloadcloneembedreportprint

0.76

 Explicit

0.73

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

0.71

actionDate

0.70

Activations Density 0.018%

user ratings followed by numbers

elements related to ratings and reviews

No Comments

No Known Activations