INDEX

Explanations

dialogue marks and interviewer names in a conversation

oai_token-act-pair · gpt-3.5-turbo

instances of emotional expressions and personal sentiments

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

 undermin

-0.93

 proport

-0.88

 occas

-0.86

 exha

-0.81

 eleph

-0.77

aditional

-0.76

 citiz

-0.76

 tremend

-0.76

everal

-0.74

 footing

-0.73

POSITIVE LOGITS

AMY

0.97

↵

0.85

GA

0.81

NER

0.81

Laughs

0.81

SIM

0.81

Jerry

0.79

SI

0.79

Interview

0.79

GREEN

0.77

Activations Density 0.366%

dialogue marks and interviewer names in a conversation

instances of emotional expressions and personal sentiments

No Comments

No Known Activations