INDEX

Explanations

phrases related to investigative actions or digging into information

oai_token-act-pair · gpt-3.5-turbo

phrases that involve exploration or investigation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

uploads

-0.87

rican

-0.72

roy

-0.68

benefit

-0.65

dom

-0.64

sup

-0.64

roit

-0.62

dominated

-0.62

owl

-0.62

SON

-0.62

POSITIVE LOGITS

 depths

0.79

 Telescope

0.76

 recess

0.76

 deeper

0.73

 Travels

0.73

 archives

0.72

 graves

0.72

 Explore

0.69

itud

0.68

 digs

0.68

Activations Density 0.123%

phrases related to investigative actions or digging into information

phrases that involve exploration or investigation

No Comments

No Known Activations