INDEX

Explanations

references to the television network "FOX"

oai_token-act-pair · gpt-3.5-turbo

mentions of the FOX network and its related content

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

acca

-0.84

ainer

-0.76

ucket

-0.75

ivari

-0.75

ylum

-0.73

adding

-0.71

asket

-0.70

anks

-0.70

ances

-0.67

 Mellon

-0.67

POSITIVE LOGITS

FOX

1.06

 SPORTS

0.93

conn

0.85

CAST

0.84

TAG

0.79

MAT

0.78

IUM

0.76

çīĪ

0.76

 Broadcast

0.76

POS

0.76

Activations Density 0.007%

references to the television network "FOX"

mentions of the FOX network and its related content

No Comments

No Known Activations