INDEX

Explanations

words related to geopolitical situations and events, particularly those involving the name "Gonski"

oai_token-act-pair · gpt-3.5-turbo

the word "ki" in various contexts

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

nard

-0.69

rique

-0.68

icals

-0.68

shire

-0.67

%]

-0.65

icist

-0.65

ication

-0.64

mort

-0.63

 Leban

-0.61

icity

-0.61

POSITIVE LOGITS

wi

1.18

pper

1.13

hao

1.04

yo

0.98

utenant

0.98

pping

0.95

yah

0.91

ota

0.90

oji

0.90

Äĩ

0.89

Activations Density 0.016%

words related to geopolitical situations and events, particularly those involving the name "Gonski"

the word "ki" in various contexts

No Comments

No Known Activations