INDEX

Explanations

instances where something is falsely claimed or accused

oai_token-act-pair · gpt-3.5-turbo

terms related to false accusations and wrongful actions

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

 willingness

-0.72

 impossibility

-0.72

 inaction

-0.71

 enthus

-0.70

 spont

-0.70

iquette

-0.68

Cra

-0.68

 openness

-0.67

combe

-0.67

 absurdity

-0.66

POSITIVE LOGITS

 priced

0.95

 portrayed

0.94

 represented

0.94

 positioned

0.90

 diagnosed

0.89

 transported

0.89

 handled

0.89

 assessed

0.88

 punished

0.88

 evaluated

0.86

Activations Density 0.152%

instances where something is falsely claimed or accused

terms related to false accusations and wrongful actions

No Comments

No Known Activations