INDEX

Explanations

estimates or measurements about quantities, often expressed in terms of "percentage points"

oai_token-act-pair · gpt-3.5-turbo

references to statistical points or percentages in data reporting

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

ufact

-0.77

lict

-0.75

 tremend

-0.67

uthor

-0.67

RAW

-0.65

 garg

-0.65

 facult

-0.63

ourt

-0.63

 notor

-0.63

rodu

-0.62

POSITIVE LOGITS

point

0.84

lash

0.82

hift

0.76

points

0.75

manship

0.74

 points

0.70

(+

0.69

uing

0.69

endment

0.68

.","

0.67

Activations Density 0.023%

estimates or measurements about quantities, often expressed in terms of "percentage points"

references to statistical points or percentages in data reporting

No Comments

No Known Activations