INDEX

Explanations

numbers related to savings or costs, especially associated with government spending or financial estimates

oai_token-act-pair · gpt-3.5-turbo

financial figures and savings related to expenditure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ese

-0.79

 Lists

-0.73

 Prophe

-0.63

oys

-0.62

 prelim

-0.62

oplan

-0.62

 Conspiracy

-0.60

venge

-0.60

 Hail

-0.60

 preceded

-0.59

POSITIVE LOGITS

-$

0.86

 savings

0.82

 wasted

0.79

 lost

0.79

profits

0.78

Increase

0.77

$.

0.77

efficiency

0.75

gain

0.75

 annually

0.74

Activations Density 0.355%

numbers related to savings or costs, especially associated with government spending or financial estimates

financial figures and savings related to expenditure

No Comments

No Known Activations