INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to locations or places

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

rd

-0.75

arb

-0.75

agascar

-0.70

unes

-0.67

CHAT

-0.66

olyn

-0.65

onday

-0.64

umin

-0.63

icial

-0.63

luaj

-0.63

POSITIVE LOGITS

holder

1.17

holders

1.13

bos

0.99

upon

0.81

ername

0.77

liness

0.75

knife

0.74

 holder

0.74

bo

0.73

ngth

0.72

Activations Density 0.043%

locations or places

references to locations or places

No Comments

No Known Activations