INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to geographic or specific locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

deen

-0.77

graduate

-0.72

 Thousand

-0.72

ħĭ

-0.71

RAW

-0.71

ewater

-0.70

vernment

-0.69

lesh

-0.68

emb

-0.68

 Dying

-0.67

POSITIVE LOGITS

loc

1.29

ator

0.92

ational

0.85

ators

0.84

ãĥŁ

0.82

inia

0.80

itud

0.79

ãĤ¼ãĤ¦ãĤ¹

0.79

 mosquit

0.78

ãĤ¤ãĥĪ

0.76

Activations Density 0.016%

locations or places

references to geographic or specific locations

No Comments

No Known Activations