INDEX

Explanations

phrases related to inclusion or incorporation

oai_token-act-pair · gpt-3.5-turbo

phrases that introduce lists or specific categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ossal

-0.82

okane

-0.81

ogun

-0.81

oire

-0.77

psons

-0.76

ride

-0.76

izens

-0.76

oled

-0.75

olitics

-0.73

ifest

-0.71

POSITIVE LOGITS

lihood

0.80

 lowering

0.80

 ensuring

0.79

 preventing

0.79

 removing

0.78

 avoiding

0.77

 aesthetics

0.76

 pacing

0.74

:-

0.74

 fingerprints

0.73

Activations Density 0.155%

phrases related to inclusion or incorporation

phrases that introduce lists or specific categories

No Comments

No Known Activations