INDEX

Explanations

assertive statements implying suggestions or directives

oai_token-act-pair · gpt-3.5-turbo

repeated uses of the word "should"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Fra

-0.71

AME

-0.69

quickShipAvailable

-0.69

 Hilbert

-0.67

Dia

-0.63

anka

-0.60

Rox

-0.60

dup

-0.60

atile

-0.60

 Byrd

-0.60

POSITIVE LOGITS

 beware

1.19

nt

1.14

be

1.10

 strive

1.09

 ideally

1.07

ered

1.06

 aspire

1.05

 definitely

1.02

 reconsider

1.00

 consider

1.00

Activations Density 0.084%

assertive statements implying suggestions or directives

repeated uses of the word "should"

No Comments

No Known Activations