INDEX

Explanations

comparisons or descriptions related to strength or growth

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "stronger."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

rolley

-0.82

 Sloan

-0.75

ffe

-0.75

pty

-0.73

 externalToEVAOnly

-0.70

rection

-0.70

 Roche

-0.68

ette

-0.67

chal

-0.67

 Correction

-0.66

POSITIVE LOGITS

 strength

0.89

nesses

0.85

 promot

0.84

bodied

0.81

 motiv

0.77

 coupling

0.77

 gust

0.75

 encryption

0.75

 advoc

0.74

 vibrations

0.74

Activations Density 0.017%

comparisons or descriptions related to strength or growth

repeated mentions of the word "stronger."

No Comments

No Known Activations