INDEX

Explanations

mentions of credibility and undermining actions or entities

oai_token-act-pair · gpt-3.5-turbo

references to credibility and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Shop

-0.77

avy

-0.72

opa

-0.71

xon

-0.70

sets

-0.69

hib

-0.68

 Sieg

-0.67

EMS

-0.67

allow

-0.65

cise

-0.65

POSITIVE LOGITS

ibly

1.05

 credibility

1.00

 legitimacy

0.99

acies

0.89

 credentials

0.87

worthiness

0.78

 tremend

0.77

itimate

0.77

 validity

0.76

 guiActiveUn

0.76

Activations Density 0.015%

mentions of credibility and undermining actions or entities

references to credibility and related concepts

No Comments

No Known Activations