INDEX

Explanations

mentions of scientific research being published in journals

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "published."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

llan

-0.90

hart

-0.77

xa

-0.71

vette

-0.68

 Architects

-0.68

Ĭ±

-0.67

 awakened

-0.66

aturation

-0.65

uth

-0.65

ichael

-0.65

POSITIVE LOGITS

lishing

0.99

 excerpts

0.93

lisher

0.92

lishes

0.75

Ô

0.70

 behavi

0.70

itatively

0.69

 exploits

0.69

URL

0.69

 newsp

0.68

Activations Density 0.027%

mentions of scientific research being published in journals

occurrences of the word "published."

No Comments

No Known Activations