INDEX

Explanations

words related to physical harm or attack

oai_token-act-pair · gpt-3.5-turbo

instances of certain names and terms related to entities or characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

omen

-0.80

athlon

-0.76

ldon

-0.75

ples

-0.75

 handshake

-0.73

etermination

-0.73

ally

-0.73

emet

-0.72

icrobial

-0.72

EStreamFrame

-0.72

POSITIVE LOGITS

 Canaver

0.82

 Beng

0.74

glers

0.74

 Sebast

0.74

ABE

0.73

 Kuala

0.70

 Sebastian

0.69

 Pengu

0.65

Footnote

0.65

Ops

0.64

Activations Density 0.033%

words related to physical harm or attack

instances of certain names and terms related to entities or characters

No Comments

No Known Activations