INDEX

Explanations

subtle patterns of data assignment or allocation

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "assigned" and its various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

FORE

-0.82

Ïī

-0.76

mph

-0.75

bite

-0.72

bury

-0.72

FACE

-0.71

Ste

-0.71

razil

-0.71

sonian

-0.70

steen

-0.70

POSITIVE LOGITS

 assigned

0.93

 assigning

0.89

 assign

0.83

essor

0.82

 assignment

0.80

 assigns

0.77

 assignments

0.76

 attendant

0.76

istant

0.75

itory

0.74

Activations Density 0.011%

subtle patterns of data assignment or allocation

occurrences of the word "assigned" and its various contexts

No Comments

No Known Activations