INDEX

Explanations

words related to research findings or evidence

oai_token-act-pair · gpt-3.5-turbo

instances of the word "shown" associated with evidence or findings

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

izzy

-0.65

raft

-0.61

politics

-0.60

law

-0.60

draft

-0.60

dar

-0.59

ictionary

-0.58

rine

-0.57

ire

-0.56

polit

-0.55

POSITIVE LOGITS

 shown

3.62

shown

2.35

 displayed

1.89

 depicted

1.84

 demonstrated

1.77

 pictured

1.69

 seen

1.63

 exhibited

1.62

 showed

1.57

 illustrated

1.53

Activations Density 0.015%

words related to research findings or evidence

instances of the word "shown" associated with evidence or findings

No Comments

No Known Activations