INDEX

Explanations

phrases related to the foundation or origin of concepts or ideologies

oai_token-act-pair · gpt-3.5-turbo

references to underlying foundations or justifications for beliefs and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ura

-0.79

lead

-0.70

fax

-0.70

 Appears

-0.69

ricular

-0.68

mouth

-0.67

atel

-0.67

finals

-0.66

unic

-0.66

emy

-0.66

POSITIVE LOGITS

 notions

1.02

 misconceptions

0.97

 perceptions

0.96

 assumptions

0.95

 considerations

0.94

 misunderstanding

0.94

 fears

0.93

 misunderstand

0.91

 recognizing

0.90

 resentment

0.88

Activations Density 0.215%

phrases related to the foundation or origin of concepts or ideologies

references to underlying foundations or justifications for beliefs and actions

No Comments

No Known Activations