INDEX

Explanations

phrases related to debates, controversies, investigations, and questioning arguments or situations

oai_token-act-pair · gpt-3.5-turbo

references to topics being debated or questioned

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

peak

-0.82

tumblr

-0.76

âĿ

-0.75

hops

-0.73

cil

-0.73

TED

-0.71

sted

-0.71

agree

-0.69

heat

-0.67

SPONSORED

-0.67

POSITIVE LOGITS

 legality

1.63

 validity

1.60

 feasibility

1.52

 adequ

1.48

 merits

1.42

 extent

1.42

 legitimacy

1.40

 viability

1.40

 motivations

1.40

 motives

1.39

Activations Density 0.241%

phrases related to debates, controversies, investigations, and questioning arguments or situations

references to topics being debated or questioned

No Comments

No Known Activations