INDEX

Explanations

names of a person or brand "Strong"

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual associated with a significant event

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 externalToEVAOnly

-0.83

mits

-0.74

MIA

-0.72

ktop

-0.71

UTE

-0.67

RECT

-0.66

opus

-0.63

 McCann

-0.63

oreal

-0.62

çĦ

-0.62

POSITIVE LOGITS

strong

1.16

enger

1.01

itudinal

0.92

est

0.92

nesses

0.88

er

0.85

fast

0.82

Weak

0.80

itude

0.79

GoldMagikarp

0.76

Activations Density 0.008%

names of a person or brand "Strong"

references to a specific individual associated with a significant event

No Comments

No Known Activations