INDEX

Explanations

references to aquatic mammals, specifically dolphins

oai_token-act-pair · gpt-3.5-turbo

references to dolphins

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

ablishment

-0.82

rade

-0.76

eners

-0.75

rooms

-0.72

raine

-0.71

Interstitial

-0.69

cation

-0.66

ãĥ´ãĤ¡

-0.65

 Flavoring

-0.65

APD

-0.64

POSITIVE LOGITS

 dolphin

0.88

arium

0.87

 dolphins

0.79

iform

0.78

 Swim

0.77

bone

0.72

ragon

0.71

 Seal

0.71

ocamp

0.68

 seals

0.67

Activations Density 0.014%

references to aquatic mammals, specifically dolphins

references to dolphins

No Comments

No Known Activations