INDEX

Explanations

percentages and numerical proportions

oai_token-act-pair · gpt-3.5-turbo

phrases related to sources of funding and financial transactions

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

pedia

-0.65

 throne

-0.62

 crus

-0.62

sonian

-0.60

aci

-0.60

 Relations

-0.58

 arose

-0.57

 dictates

-0.56

 hither

-0.56

worthiness

-0.56

POSITIVE LOGITS

Ò

0.70

 unexpl

0.69

 concentrated

0.68

ãĥĥãĥī

0.66

 attributable

0.66

ECD

0.66

 untouched

0.65

 misdem

0.65

 unchanged

0.65

Mit

0.64

Activations Density 0.449%

percentages and numerical proportions

phrases related to sources of funding and financial transactions

No Comments

No Known Activations