INDEX

Explanations

names of specific island locations

oai_token-act-pair · gpt-3.5-turbo

references to various islands

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

apers

-0.83

ivated

-0.77

baugh

-0.75

ERA

-0.75

IENCE

-0.73

hetically

-0.72

ENG

-0.71

 Hoover

-0.71

ISON

-0.70

hetical

-0.70

POSITIVE LOGITS

 Islands

1.38

 islands

0.85

olid

0.76

hips

0.76

hell

0.74

 Borders

0.73

ylum

0.73

 Templar

0.72

ktop

0.71

conservancy

0.70

Activations Density 0.031%

names of specific island locations

references to various islands

No Comments

No Known Activations