INDEX

Explanations

instances where a specific action or instruction is mentioned or required

oai_token-act-pair · gpt-3.5-turbo

the word "specify" and its variations in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Pear

-0.70

java

-0.69

ears

-0.68

bill

-0.66

study

-0.62

Jol

-0.61

aea

-0.61

be

-0.61

cam

-0.60

lymp

-0.60

POSITIVE LOGITS

 specify

1.00

 specified

0.86

 explicitly

0.84

 specifies

0.81

 exempted

0.80

specified

0.79

wcsstore

0.78

 labels

0.78

ministic

0.78

encies

0.76

Activations Density 0.018%

instances where a specific action or instruction is mentioned or required

the word "specify" and its variations in various contexts

No Comments

No Known Activations