INDEX

Explanations

words related to selling or trading, particularly "brokers."

oai_token-act-pair · gpt-3.5-turbo

references to brokers, particularly in a transactional or investigative context

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

 align

-0.73

 reciproc

-0.63

mat

-0.61

 status

-0.61

 bear

-0.58

 content

-0.57

 level

-0.57

 vest

-0.57

 ancestor

-0.56

 required

-0.56

POSITIVE LOGITS

kers

4.66

ker

2.98

king

1.96

kies

1.91

ked

1.74

kes

1.56

kens

1.45

KER

1.36

kas

1.35

kie

1.32

Activations Density 0.008%

words related to selling or trading, particularly "brokers."

references to brokers, particularly in a transactional or investigative context

No Comments

No Known Activations