INDEX

Explanations

words related to explosive events or situations

oai_token-act-pair · gpt-3.5-turbo

references to a specific entity or subject associated with 'Bomb' and 'Nim.'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 disse

-0.77

ceiver

-0.69

 preacher

-0.69

pires

-0.67

 compassionate

-0.67

meta

-0.65

facing

-0.65

 athe

-0.64

essions

-0.63

LDS

-0.63

POSITIVE LOGITS

 Bomb

3.29

Bomb

2.50

bomb

1.69

 Bomber

1.47

 Saturn

1.41

 Bombs

1.39

 Boll

1.24

Bom

1.16

Nim

1.09

 bomber

1.05

Activations Density 0.025%

words related to explosive events or situations

references to a specific entity or subject associated with 'Bomb' and 'Nim.'

No Comments

No Known Activations