INDEX

Explanations

terms related to quantities, such as numbers, counts, or measures

oai_token-act-pair · gpt-3.5-turbo

terms related to quantities and statistics

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

cffffcc

-0.85

ĪĴ

-0.83

ãĤ´ãĥ³

-0.66

ivot

-0.66

hement

-0.65

enture

-0.65

ĺħ

-0.64

utics

-0.63

reck

-0.60

odcast

-0.59

POSITIVE LOGITS

 involved

0.81

 apiece

0.79

 available

0.78

 needed

0.77

 required

0.76

per

0.75

 allocated

0.73

 consumed

0.71

involved

0.71

izable

0.69

Activations Density 0.259%

terms related to quantities, such as numbers, counts, or measures

terms related to quantities and statistics

No Comments

No Known Activations